VMs loosing network connectivity for a few minutes on a daily basis

Hello,

I am testing VMM/VMD on OpenBSD 6.4 with OpenBSD 6.4 virtual machines but noticed that maybe around 2 times per day the VM loose their network connectivity for a short amount of time of around 2-3 minutes. I currently have 3 OpenBSD VM with very light load on them and it happens to all of them.

The network connectivity recovers on its own or if I login through the console to the VM and initiate for example a ping to the outside. The host/hypervisor itself never looses connectivity.

Now I presume there is either an issue with my network setup or maybe a bug but I would rather think it has to do with my network setup. My network setup on the OpenBSD host itself consists of two physical network devices (bnx0 + bnx1) which I have bundled in a trunk (trunk0) in failover mode. Then on top of my trunk I have two VLAN interfaces (vlan2 and vlan6). vlan2 is my private network and vlan6 is my public facing network (internet). Then finally I have a bridge interface (bridge6) with my vlan6 interface inside where my VM connect to as they are directly available on the internet.

So the whole chain of network interfaces from host to VM looks like this:

[bnx0+bnx1]-[trunk0]-[vlan6]-[bridge6]-[tap0]-[vio0]


My /etc/vm.conf looks like this:

switch "uplink_vlan6" {
interface bridge6
}

vm "obsd1vm" {
memory 2G
disk "/var/vmm/obsd1vm.qcow2"

interface {
switch "uplink_vlan6"
lladdr fe:e1:bb:03:01:01
}
}

My /etc/hostname.* files look like this:

/etc/hostname.bnx0
up

/etc/hostname.bnx1
up

/etc/hostname.trunk0
trunkproto failover trunkport bnx0 trunkport bnx1 up

/etc/hostname.vlan2
inet 192.168.1.56 255.255.255.0 192.168.1.255 vnetid 2 parent trunk0 description "private" up

/etc/hostname.vlan6
inet xxx.xxx.xxx.xxx 255.255.255.0 xxx.xxx.xxx.255 vnetid 6 parent trunk0 description "public" up

/etc/hostname.bridge6
add vlan6
up

The hardware switch behind the host is a Cisco switch and the two ports connected to the two hardware NICS of the server have both the following config:

interface Eth101/1/9
switchport mode trunk
switchport trunk native vlan 99
switchport trunk allowed vlan 2,6


Finally below is the output of ifconfig:

bnx0: flags=8b43<UP,BROADCAST,RUNNING,PROMISC,ALLMULTI,SIMPLEX,MULTICAST> mtu 1500
lladdr ---REMOVED---
index 3 priority 0 llprio 3
trunk: trunkdev trunk0
media: Ethernet autoselect (1000baseT full-duplex,rxpause)
status: active
bnx1: flags=8b43<UP,BROADCAST,RUNNING,PROMISC,ALLMULTI,SIMPLEX,MULTICAST> mtu 1500
lladdr ---REMOVED---
index 4 priority 0 llprio 3
trunk: trunkdev trunk0
media: Ethernet autoselect (1000baseT full-duplex,rxpause)
status: active
bridge6: flags=41<UP,RUNNING>
description: switch1-uplink_vlan6
index 5 llprio 3
groups: bridge
priority 32768 hellotime 2 fwddelay 15 maxage 20 holdcnt 6 proto rstp
vlan6 flags=3<LEARNING,DISCOVER>
port 8 ifpriority 0 ifcost 0
tap0 flags=3<LEARNING,DISCOVER>
port 10 ifpriority 0 ifcost 0
trunk0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
lladdr ---REMOVED---
index 6 priority 0 llprio 3
trunk: trunkproto failover
trunkport bnx1
trunkport bnx0 master,active
groups: trunk
media: Ethernet autoselect
status: active
vlan2: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
lladdr ---REMOVED---
description: private
index 7 priority 0 llprio 3
encap: vnetid 2 parent trunk0
groups: vlan egress
media: Ethernet autoselect
status: active
inet 192.168.1.56 netmask 0xffffff00 broadcast 192.168.1.255
vlan6: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
lladdr ---REMOVED---
description: public
index 8 priority 0 llprio 3
encap: vnetid 6 parent trunk0
groups: vlan
media: Ethernet autoselect
status: active
inet ---REMOVED--- netmask 0xffffff00 broadcast ---REMOVED---
pflog0: flags=141<UP,RUNNING,PROMISC> mtu 33136
index 9 priority 0 llprio 3
groups: pflog
tap0: flags=8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
lladdr fe:e1:ba:d0:56:1c
description: vm1-if0-obsd1vm
index 10 priority 0 llprio 3
groups: tap
status: active

Last note, the host and VMs are all patched up to 013_unveil.

I hope I could provide here all the relevant details, if there is anything else I should add I would be happy to provide with more info.

Best regards,
Mabi